Showing papers by "Barry O'Sullivan published in 2007"

Representative explanations for over-constrained problems

Abstract: In many interactive decision making scenarios there is often no solution that satisfies all of the user's preferences. The decision process can be helped by providing explanations. Relaxation show sets of consistent preferences and, thus, indicate Which preferences can be enforced, while exclusion sets show which preferences can be relaxed to obtain a solution. We propose a new approach to explanation based on the notion of a representative set of explanations. The size of the set of explanations we compute is exponentially more compact than that found using common approaches from the literature based on finding all minimal conflicts.

Query-driven constraint acquisition

Abstract: The modelling and reformulation of constraint networks are recognised as important problems. The task of automatically acquiring a constraint network formulation of a problem from a subset of its solutions and non-solutions has been presented in the literature. However, the choice of such a subset was assumed to be made independently of the acquisition process. We present an approach in which an interactive acquisition system actively selects a good set of examples. We show that the number of examples required to acquire a constraint network is significantly reduced using our approach.

Trends in Constraint Programming

Abstract: Introduction. Part I. The Past, Present and Future of Constraint Programming. Chapter 1. Constraint Programming as Declarative Algorithmics. Chapter 2. Constraint Programming Tools. Chapter 3. The Next 10 Years of Constraint Programming. Chapter 4. Constraint Propagation and Implementation. Chapter 5. On the First SAT/CP Integration Workshop. Chapter 6. Constraint-based Methods for Bioinformatics. Part II. Constraint Modeling and Reformulation. Chapter 7. Improved Models and Reformulation. Chapter 8. The Automatic Generation of Redundant Representations and Channeling Constraints. Part III. Symmetry in Constraint Satisfaction Problems. Chapter 9. GAPLex: Generalized Static Symmetry Breaking. Chapter 10. Symmetry Breaking in Subgraph Pattern Matching. Part IV. Interval Analysis, Constraint Propagation and Applications. Chapter 11. Modeling and Solving of a Radio Antenna Deployment Support Application. Chapter 12. Guaranteed Numerical Injectivity Test via Interval Analysis. Chapter 13. An Interval-based Approximation Method for Discrete Changes in Hybrid cc. Part V. Local Search Techniques in Constraint Satisfaction. Chapter 14. Combining Adaptive Noise and Look-Ahead in Local Search for SAT. Chapter 15. Finding Large Cliques using SAT Local Search. Chapter 16. Multi-Point Constructive Search for Constraint Satisfaction: An Overview. Chapter 17. Boosting SLS Using Resolution. Chapter 18. Growing COMET. Part VI. Preferences and Soft Constraints. Chapter 19. The Logic Behind Weighted CSP. Chapter 20. Dynamic Heuristics for Branch and Bound on Tree-Decomposition of Weighted CSPs. Part VII. Constraints in Software Testing, Verification and Analysis. Chapter 21. Extending a CP Solver with Congruences as Domains for Program Verification. Chapter 22. Generating Random Values Using Binary Decision Diagrams and Convex Polyhedra. Chapter 23. A Symbolic Model for Hash-Collision Attacks. Chapter 24. Strategy for Flaw Detection Based on a Service-driven Model for Group Protocols. Part VIII. Constraint Programming for Graphical Applications. Chapter 25. Trends and Issues in using Constraint Programming for Graphical Applications. Chapter 26. A Constraint Satisfaction Framework for Visual Problem Solving. Chapter 27. Computer Graphics and Constraint Solving: An Application to Virtual Camera Control. Index.

Distance constraints in constraint satisfaction

Abstract: Users can often naturally express their preferences in terms of ideal or non-ideal solutions. We show how to reason about logical combinations of distance constraints on ideals and non-ideals using a novel global constraint. We evaluate our approach on both randomly generated and real-world configuration problem instances.

Quantified constraint satisfaction problems: from relaxations to explanations

Abstract: The Quantified Constraint Satisfaction Problem (QCSP) is a generalisation of the classical CSP in which some of variables can be universally quantified. In this paper, we extend two well-known concepts in classical constraint satisfaction to the quantified case: problem relaxation and explanation of inconsistency. We show that the generality of the QCSP allows for a number of different forms of relaxation not available in classical CSP. We further present an algorithmfor computing a generalisation of conflict-based explanations of inconsistency for the QCSP.

Truthful risk-managed combinatorial auctions

Abstract: Given a winning-bid withdrawal in a combinatorial auction, finding an alternative repair solution of adequate revenue without causing undue disturbance to the remaining winning bids in the original solution may be difficult or even impossible. This "bid-takers exposure problem" may be preemptively addressed by finding a solution that is robust to winning-bid withdrawal. We introduce the concept of monotonicity-in-expectation. We provide impossibility results concerning truthful mechanisms for robust solutions with bounded social-welfare losses in which the bid-taker cannot rescind items from winning bidders to repair a solution. We also show that this result extends to combinatorial auctions that include a form of leveled-commitment contract. However, we present a positive result regarding truthfulness for combinatorial auctions in a restricted setting that comprises a computationally efficient allocation algorithm that seeks to maximize expected social welfare.

Semiring-Based Constraint Acquisition

Abstract: Constraint programming offers a declarative approach to solving problems modeled as constraint satisfaction problems (CSPs). However, the precise specification of a set of constraints is sometimes not available, but may have to be learned, for instance, from a set of examples of its solutions and non-solutions. In general, one may wish to learn generalized CSPs involving classical, fuzzy, weighted or probabilistic constraints, for example. This paper introduces a unifying framework for CSP learning. The framework is generic in that it can be instantiated to obtain specific formulations for learning classical, fuzzy, weighted or probabilistic CSPs. In particular, a new formulation for classical CSP learning, which minimizes the number of examples violated by candidate CSPs, is obtained by instantiating the framework. This formulation is equivalent to a simple pseudo-boolean optimization problem, thus being efficiently solvable using many optimization tools.

Directed Feedback Vertex Set is Fixed-Parameter Tractable

Abstract: We resolve positively a long standing open question regarding the fixed-parameter tractability of the parameterized Directed Feedback Vertex Set problem. In particular, we propose an algorithm which solves this problem in $O(8^kk!*poly(n))$.

Constraint symmetry for the soft CSP

Abstract: We introduce a definition of constraint symmetry for soft CSPs, based on the definition of constraint symmetry for classical CSPs. We show that the constraint symmetry group of a soft CSP is a subgroup of that of an associated classical CSP instance. Where it is smaller, we can successfully exploit the additional symmetries using conditional symmetry breaking. We demonstrate, in a case-study of graph colouring, that eliminating the symmetry of the soft CSP combined with conditional symmetry breaking can lead to huge reductions in the search effort to find an optimal solution to the soft CSP.

Generating and solving logic puzzles through constraint satisfaction

Abstract: Solving logic puzzles has become a very popular past-time, particularly since the Sudoku puzzle started appearing in newspapers all over the world. We have developed a puzzle generator for a modification of Sudoku, called Jidoku, in which clues are binary disequalities between cells on a 9 × 9 grid. Our generator guarantees that puzzles have unique solutions, have graded difficulty, and can be solved using inference alone. This demonstration provides a fun application of many standard constraint satisfaction techniques, such as problem formulation, global constraints, search and inference. It is ideal as both an education and outreach tool. Our demonstration will allow people to generate and interactively solve puzzles of user-selected difficulty, with the aid of hints if required, through a specifically built Java applet.

Directed Feedback Vertex Set is Fixed-Parameter Tractable

Abstract: We resolve positively a long standing open question regard- ing the fixed-parameter tractability of the parameterized Directed Feed- back Vertex Set problem. In particular, we propose an algorithm which solves this problem in O(8 k k! poly(n)).

Generalized constraint acquisition

Abstract: Constraint programming is an approach to problem solving that relies on a combination of inference and search to solve real-world problems formulated as constraint satisfaction problems (CSPs). Many methods for solving CSPs have been developed. However, the specification of a CSP is sometimes not available, but may have to be learned from a training set, which is given, for instance, as a set of examples of its solutions and non-solutions. The motivating applications for constraint acquisition are many. For example, often one may wish to find a compact representation of a CSP instance for purposes such as explanation generation, requirements gathering, and specification. Acquiring soft constraints, which we focus on here, can be regarded as learning about preferences, uncertainty or costs in a combinatorial setting.

Search Ordering Heuristics for Restarts-based Constraint Solving

Abstract: Over the past decade impressive advances have been made in solving Constraint Satisfaction Problems by using of randomization and restarts. In this paper we propose a new class of variable and value ordering heuristics based on learning from nogoods without storing them. We show empirically that these heuristics dramatically improve the performance of restarts-based constraint solving.

A reformulation-based approach to explanation in constraint satisfaction

Abstract: Many approaches to explanation generation have been reported in the literature [1-5]. The dominant approach to explanation is based on computing minimal conflicting sets of constraints. An explanation can be considered concise if it involves few constraints of low arity. However, in many practical domains constraints are specified extensionally as table constraints. From an explanation point of view, table constraints can prevent us from finding concise explanations since the constraint may be specified over all the variables in the problem. However, it is often possible to reformulate a large arity table constraint into a set of low arity constraints whose join is logically equivalent to the original table constraint. In this research we are concerned with finding reformulations of large arity table constraints into low arity constraints in order to help find more concise explanations in interactive applications such as product configuration or online electronic commerce.